High-Throughput Monoclonal Antibody Discovery from Phage Libraries: Challenging the Current Preclinical Pipeline to Keep the Pace with the Increasing mAb Demand

Abstract

Monoclonal antibodies are among the most powerful therapeutics in modern medicine. Since the approval of the first therapeutic antibody in 1986, monoclonal antibodies keep holding great expectations for application in a range of clinical indications, highlighting the need to provide timely and sustainable access to powerful screening options. However, their application in the past has been limited by time-consuming and expensive steps of discovery and production. The screening of antibody repertoires is a laborious step; however, the implementation of next-generation sequencing-guided screening of single-chain antibody fragments has now largely overcome this issue. This review provides a detailed overview of the current strategies for the identification of monoclonal antibodies from phage display-based libraries. We also discuss the challenges and the possible solutions to improve the limiting selection and screening steps, in order to keep pace with the increasing demand for monoclonal antibodies.

Keywords: high-throughput screening; mAbs production; monoclonal antibodies; next-generation sequencing; selection-related; selection-unrelated; target-unrelated; third-generation sequencing.

Figure 1

The cartoon depicts the entire procedure of in vitro biopanning. The biopanning of phage display libraries involves three principal steps for selection. ① A negative selection is performed by incubating the scFv library (and sub-libraries) with all the components of the screening (in the example Fc-Tag and solid-phase) except for the antigen of interest to discard target-unrelated clones. In the positive selection step, phages are incubated with the antigen of interest ②; unbound phages are washed out by extensive washes. Residual target-unrelated clones are indicated in the zoomed circle as a binder to Fc-Tag. Bound phages from positive selection are eluted ③ and are amplified in E. coli by co-infection with helper phage ④. All these steps are repeated several times to enrich the target-specific clones.

Figure 2

Strategies to analyze phage library by second- and third-generation sequencing. On the left, the workflow for extraction of variable fragments compatible with second-generation sequencing platforms (e.g., Illumina MiSeq) is shown. Full-length scFvs and VH fragments are extracted from dsDNA of phages from E. coli. On the right, the workflow for the application of third-generation sequencing of entire scFvs by circular consensus sequencing is represented. Molecular barcoding can be included in both strategies for multiplexing samples. The indicated pros (green font) and cons (red font) are referred to as the current state-of-the-art.

Figure 3

Funnel pipeline applied to sequencing data to improve the confidence of in silico binding predictions.

Figure 4

Snapshot of sub-libraries composition in different scFv phage display screenings. Potential binders, clones bearing stop codons or altered ORFs, and target-unrelated clones are respectively shown in light green, red and orange. Dark green indicates the percentage of specific binders validated by ELISA assays, where available. Ref relative to the screening: CTLA-4 [47]; PD1 to TIM3 [33]; CLDN-1 [61].

Figure 5

Rescue strategies to physically isolate a hypothetical clone of interest (green) from a given sub-library. (A) In the inverse PCR approach, two primers are designed into the HCDR3 of the scFv to obtain the newly synthesized, entire phagemid vector of interest. DpnI enzyme is used to cleave and remove template DNA containing methylated and hemi-methylated restriction sites (G^mA|TC) preserving the newly synthesized, unmethylated DNA. (B) In overlapping PCR, primers into the HCDR3 region are combined with primers designed into constant HFR1 and LFR4. The two amplicons are assembled by overlapping PCR and cloned into an empty phagemid vector. (C) In hybridization capture, functionalized primers specific to HCDR3 are used as bait to isolate the clone of interest.

Figure 6

Workflow for scFvs conversion into full mAbs and small/medium scale production. (A) Variable heavy and light chains of an scFv can be subcloned into mammalian expression vectors encoding constant domains of antibodies. A toxic gene can be used to improve the confidence of cloning. Any format or isotype of interest can be produced (full mAb, scFv-Fc, etc.). (B) Small scale (e.g., 12-well plates) transient/semi-stable transfections in the producer cell lines allow for obtaining enough antibodies for preliminary in vitro testing. The transfection/purification process can be scaled up for preclinical testing.

Figure 7

Long non-coding SINEUP RNAs can be used to improve the production of recombinant antibodies (full mAbs or scFv-Fc) by divergent 5′ end head-to-head complementarity with the target coding mRNA. The binding domain of SINEUP has been designed to target the region encompassing 5′UTR and signal peptide of any secreted recombinant protein.